harisrid Tech News

“When a Measure Becomes a Target, It Ceases to be a Good Measure” – British Economist Charles Goodhart’s

Correlation is not causation

I’ve had to work with a couple of Enterprise platforms in the past – as well as do some studying or reading of other blog posts on the side. Even though my day-to-day role is software engineering, I’ve encountered the topics tangentially via 1:1 chats with product managers or in my study of designing systems. In doing so, I’ve observed a raft number of metrics to definitely “be on the lookout for“. Let me review some of them.

These metrics matter. They matter because they help us think of how to evaluate a platform’s success. It’s easy to throw in platforms and dashboards anywhere, plop a single number, and assert “a growth in X over multiple quarters or days clearly indicates a successful product”.

Correlation is Not Causation

And metrics are far far more nuanced. Usually, the more you collect, the better ( until you run into the law of diminishing returns ).

Commonly Built-Out Enterprise Platforms:

1. Internal Data Governance, Controls, and Privacy
2. Data Management
3. Real-time advertising and ads analytics
4. High-volume end-to-end financial payments systems

A Metrics List

1. Number of personas onboarded
   • Teams
   • Customers/Stakeholders
   • End users
2. Number of use cases onboarded
   • Business workflows
   • Business use cases
3. Advertising
   • Campaign effectiveness : in totality, across segment groups and business domains
   • Profit associated per ad
4. Click rates
   • Click number reduction on webpage drill down ( e.g. 7 clicks -> 4 clicks )
   • Increased click rates on ads
   • Increased click rates on E-commerce cart buttons ( e.g. check out or add to cart )
5. Conversion rates
   • Number of customers who sign up ( per month/per year )
   • Number of customers retained after a sign up ( after a month/year )
6. User base
   • Number of users
     • Number of active users ( DAU )
7. Growth in AUM ( Assets Under Management )
   • Number of databases
   • Number of data sets
   • Financial records volume
8. Session time
   • Decreasing session length
   • Increases in lengths ( in target areas – e.g. e-Commerce catalogs )
9. Operations time
   • Decreasing time in “annoying” operations : sign-up, log in
   • Increase in rates of preferred operations : sharing of referall links
10. Page Views
    • Number of views
    • Number of persistent views ( e.g. full video length or 5-minutes into a video )
    • Number of unique views ( per visitor ) ( per source )